Milk-Marketing: Impact of Perceived Quality on Consumption Patterns
Carola Grebitus, Chengyan Yue, Maike Bruhn, Helen H. Jensen University of Kiel, Germany; Iowa State University, USA
[email protected]
Contributed Paper prepared for presentation at the 105th EAAE Seminar ‘International Marketing and International Trade of Quality Food Products’, Bologna, Italy, March 8-10, 2007
Copyright 2007 by C. Grebitus, C. Yue, M. Bruhn and H.H. Jensen. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies.
Milk-Marketing: Impact of Perceived Quality on Consumption Patterns Carola Grebitus 1, Chengyan Yue 2, Maike Bruhn 3, Helen H. Jensen 1,3 University of Kiel, Germany; 2,4 Iowa State University, USA
4
[email protected]
Summary Consumers’ use of quality characteristics to make milk purchase decisions reveal opportunities to create successful marketing strategies. Such a strategy could concern food quality. In this case, three core areas influence consumers’ quality perception: the perception process, the physical product itself and the communication about it (Grunert et al., 1996). Beyond this background, this article analyzes the impact of certain quality characteristics and socio-demographics on consumption patterns regarding whole fat milk, skim milk and organic milk. These milks were chosen because of the increasing awareness of different fat contents in the meaning of lower fat contents being healthier and the increasing importance of the organic food market. Steenkamp’s (1990) concept of ‘perceived quality’ provides as theoretical background. To gather the data a consumer survey with 260 households took place in Germany in 2004. An ordered logit model and a cluster analysis were used for analyzing the data. We find clear differences in consumers’ perception of quality characteristics for the different milks. This information can be used to develop demand-oriented marketing activities (Kotler and Armstrong, 1994: 48). KEYWORDS: milk, marketing, consumption patterns, perceived quality, ordered logit JEL: D12, Q13, M3
1.
Introduction
Today, food companies have to deal with strong competition, while consumers’ demand stagnates in many cases. The European agribusiness is characterized by saturated markets and increasingly homogeneous products. Food quality may offer one possibility for differentiation. But, any effort to differentiate products and promote food quality will only be successful if new or advanced attributes can be communicated to consumers (Alvensleben and Scheper, 1997). For consumers, certain qualities have to be visible and understandable in order to reduce uncertainty about the product and to avoid consumer dissatisfaction. To meet consumers’ expectations and preferences, it becomes important for producers and retailers to know which quality characteristics are relevant to their customers (Grunert et al., 2004). To promote food quality the following aspects have to be taken into account. Quality itself is a complex and a dynamic concept (Garvin, 1984). It refers to aspects of the food product and the basic production process that can be measured and documented in an objective way. The quality that consumers associate with a food product is often not equivalent to this objective quality evaluation (Scholderer and Bredahl, 2004). For consumers quality is a subjective concept whose association is based on psychological processes (Steenkamp, 1990). According to Cardello (1995) food quality from a consumer’s perspective is a perceptual and an evaluative construct which is related to person, place of purchase and purchase situation. Consequently in developing efficient marketing strategies the understanding of consumer quality perception is a key factor (Olson and Reynolds, 1983). Technological progress and the change of standards and norms as well as the modification of consumers’ beliefs and attitudes may cause changes in consumers’ perception (Bruhn, 2004). Technological progress and industrialisation may cause alienation and anxieties.
215
This leads to distrust of food producers, which makes it difficult to communicate credence quality attributes. Furthermore, the fact that the consumer may not be able or willing to judge experience and credence quality attributes might lead as well to an existing uncertainty about food quality (Gierl and Stumpp, 1999). In the broadest sense all these aspects influence consumer’s perception of food quality. This paper considers consumers’ perception of milk quality in particular. Thus, we want to investigate consumers’ perception process by analyzing which quality attributes influence consumption patterns taking the approach of ‘perceived quality’ (Steenkamp, 1990) into account. The products of interest for the research object are whole fat milk, skim milk and organic milk. We take into account the different fat contents as health factors. Because of the increasing importance of the organic food market, we are interested in the comparison of consumers’ quality perception of organic versus conventional foods. Since 2003 the growth rate of the German market doubled. Market data show, that Germany’s sales volume of organic products has increased considerably. It rose from € 2.05 billion in 2000 to € 3.5 billion in 2004 (ZMP, 2006b: 2). Milk was chosen as the focus of the research because it is the organic food product which is most widely distributed in Germany and is available in a great variety of shelf life, packaging, brand etc. We use an ordered logit model as a framework for identifying the quality attributes that influence consumers’ perception of quality. An application to the different milks is made in relation to consumers’ consumption frequency of these goods. Furthermore, a cluster analysis was used to segment customers according to their consumption behavior and use of the quality characteristics. Data from a household survey conducted in 2004 in Germany (n=260) allow us to relate consumer behavior to selected demographic characteristics. Finally, marketing recommendations to food producers and retailers can be made in order to promote food quality. The paper is organized as follows. The second section presents the theoretical background of the quality perception process. Section three discusses briefly the German milk consumption. In the fourth section the ordered logit model and data base is described. The fifth section displays the empirical results. Section six discusses the results and gives marketing recommendations.
2.
Background
Following Steenkamp (1990) we use the term ‘perceived quality’ to stress, that consumers’ quality evaluation is dependent on their perceptions, needs, and goals. Perceived quality is regarded as an overall one-dimensional evaluative judgment, which is based on the processing of quality cues in relation to relevant quality attributes. According to this, Steenkamp (1990) developed the model of the quality perception process on which we will base our empirical work. This model of the quality perception process describes the way consumers form perceptions about the quality of a product in purchase decisions. In the model the separation between intrinsic and extrinsic quality cues (Olson, 1978) and between experience and credence quality attributes (Nelson, 1974; Darby and Karni, 1973) is outlined. Intrinsic quality cues refer to everything of the physical product, such as color, odor, fat content. Extrinsic quality cues refer to everything else, such as point of sale, price, and brand. Only quality cues can be perceived and evaluated at the point of sale. Experience quality attributes can be evaluated after purchase or consumption, e.g. taste, and convenience. Some quality attributes can never be evaluated by the consumer him/herself. These attributes are called credence quality attributes, e.g. animal welfare, organic production. Thus, it is important to know what quality characteristics are important for consumers to make the purchase decision. If the results show, that experience or credence quality attributes influence their purchase behavior, quality cues related to those attributes would be needed for marketing campaigns.
216
The quality perception process is divided into the three sub-processes of cue acquisition and categorization, quality attribute belief formation, and integration of quality attribute beliefs, which are influenced by personal and situational variables (see figure 1). Insert figure 1 here. To categorize the quality characteristics used in the analysis we follow an approach by Caswell et al. (1998) which was expanded by Northen (2000) and Bruhn et al. (2005) to indicate the types of attribute sub-sets which exist and examples of attributes within each sub-set. In this approach the quality characteristics are divided into product, process and environmental characteristics. In the survey we took this approach to categorize the characteristics used to explain the impact of certain quality cues and attributes on consumption frequency (see table 1). Note that the examples are by no means exhaustive. Insert table 1 here.
3.
Objectives
From 2003 to 2005 Germany’s consumption of whole fat milk decreased while that of skim milk consumption increased. The average prices were relatively stable, with skim milk being cheaper than whole fat milk (see table 1). Insert table 2 here. The per capita consumption of whole fat milk and skim milk is depicted in table 3. Insert table 3 here. Regarding milk prices organic milk is more expensive than conventional milk (see table 4). Insert table 4 here. Moreover, there are differences between average prices in different food retailers. Data show that one litre whole fat milk costs 0.59 EUR at the discounter, 0.65 EUR at the hypermarket and 0.66-0.68 EUR at the supermarket (ZMP, 2002).
4.
Methodology and Data
4.1 Model Consumers’ willingness to buy (purchase intention) the three different milk products is expressed in frequency of consumption, such as daily, 5-6 times a week, 3-4 times a week, etc. to measure the corresponding latent utilities. Because the dependent variables are categorical instead of quantitative, we use an ordered logit model with robust standard errors to estimate the probability of consumers’ frequency of consumption. Suppose
U im is the utility that consumer i derives from consuming the product m and U ij
can be expressed as follows:
U im = X i β m + ε im ; i = 1,L, n ; m = 1,L M
(1)
where X i is the design matrix which is a row vector of the ith consumer’s characteristics. These characteristics include socio-demographics and quality attributes. coefficient associated with X i . And
ε
m i
β m is
the
is the residual error term that is not captured by
design matrix X i . There are n consumers and M products. In a survey that asks the respondents’ opinion, the respondents’ intensity of feelings is dependent on the measurable factors X and unobservables. In many situations, the respondents are not asked to respond to U directly. Instead, they are given only a set number of possible answers, say six, to the question of y. Consumers choose the cell that most closely represents the intensity of response to the question. For example, for product
217
m, consumer i is asked to choose among the six choices: daily ( yi = 6 ), 5-6 times a week m
( yi = 5 ), 3-4 times a week ( yi = 4 ), 1-2 times a week ( yi = 3 ), less than once a week m
m
m
( yi = 2 ), and never ( yi = 1 ). m
m
The ordered logit model depends upon the idea of the cumulative logit. This in turn relies m
on the idea of the cumulative probability. Let Cij denote the probability that the ith individual is in the jth or higher category for product m: j
Cijm = Pr ob( yim ≤ j ) = ∑ Pr ob( yim = k )
(2)
k =1
Then we turn the cumulative probability into cumulative logit for product m:
⎛ Cijm ⎞ logit(C ) = log ⎜ = α mj − β m X i ⎜ 1 − Cijm ⎟⎟ ⎝ ⎠ m ij
(3)
m=whole fate milk, skim milk, and organic milk.
4.2 Data The current survey conducted in 2004 at private households analyzes the consumption of milk. The data were collected in Germany in the capital city of the federal state SchleswigHolstein, Kiel. The randomly drawn sample consisted of 260 participants. The structure of the sample is displayed in table 5. Insert table 5 here.
5.
Results
5.1 Descriptive Results To investigate impact factors concerning the consumption patterns of conventional whole fat milk (3.5% fat), conventional skim milk (1.5% fat) and organic milk the participants were asked about their consumption frequency. They had to state how many times a week they consume the different milks. Results show that whole fat milk is preferred by 28% of the sample (see table 6). Insert table 6 here. To analyze the importance of certain quality characteristics on the consumption frequency interviewees were asked which of the following characteristics they use when making the milk purchase decision (multiple nominations). Table 7 shows the percentage of participants that decided how the given attribute is important regarding the purchase. Characteristics were categorized following Caswell et al. (1998) (see section 2). The results show that shelf-life, freshness and price are main drivers for making milk purchase decisions, while additional information, e.g. recipes, nutrition information and variety seem to have a low impact on consumers’ decisions (see table 7). Insert table 7 here. These attributes are included in the ordered logit model to show their impact on consumers’ consumption and purchase patterns.
218
5.2 Impact of Milk Quality Perception on Consumption Patterns With regard to the relation between the use of certain quality characteristics and the consumption patterns stated by the participant an ordered logit seems to be the best solution. Therefore, the attributes (table 7) are included as dummy variables in the ordered logit model to show their impact on consumers’ consumption patterns. Furthermore, the socio-demographics are independent variables. Table 8 explains the variables included in the model. We estimated three ordered logit analyses with robust standard errors with the different milks being the dependent variables. Insert table 8 here. The estimation results are reported in table 9 and 10. The rows are separated in categories of food quality characteristics and socio-demographics. The columns report the estimated coefficients, standard errors and the respective z-values of the ordered logit model explaining consumption frequency of the different milks. The results show differences for the consumers purchasing the different types of milk. We assume that consumption and purchase are correlated as 70% of the survey participants are always responsible for grocery shopping and 26% are at least sometimes responsible for grocery shopping. We will make assumptions concerning the impact of the quality characteristics on the perceived food quality and consequently on consumption and purchase. Insert table 9 here. Insert table 10 here. All three categories, product, process and environment attributes influence the milk consumption patterns. For marketing recommendations even more important is the impact of quality cues. Both extrinsic and intrinsic quality cues are significant predictors of milk consumption frequency. The extrinsic cues related to the functional product characteristic of package size increases the frequency of consumption of whole fat milk. We note that there is a lot more packaging variety for whole fat milk than for organic and skim milk. Consumers who require special sizes such as smaller packs for single households or bigger packs for families may choose whole fat milk because they have no similar alternative for organic and skim milk. The extrinsic cues related to the image product characteristic of brand increases the frequency of consumption of whole fat milk and organic milk. We note that there is a lot more brand variety for whole fat milk than for organic milk. This might be useful for communication strategies; especially, as there is no effect of brand to note for skim milk consumption. This leads to the assumption that brands of skim milk should be more often in the focus of advertisement. Furthermore, a positive effect is to state for organic milk with regard to additional information such as information about the production process or animal welfare. To provide this information could increase consumers’ benefit and thus increase their positive attitude against the product. Furthermore, the nutrition information on fat content has a significant negative effect for organic milk but a significant positive effect for skim milk. This underlines that health conscious consumers would buy skim milk. Nevertheless, concerns about fat intake may lead consumers to choose other beverages than the dairy at all. Among intrinsic quality cues ingredients discourage skim milk consumption. The effect on whole fat milk and organic milk consumption is also negative, but not statistically significant. The experience quality attribute related to sensory aspects of freshness has a significant positive effect on whole fat milk consumption. Given the perishable nature of milk, freshness would likely be an important quality attribute. Those that put high value on this attribute are more frequent consumers of whole fat milk. Regarding credence quality attributes, whole fat milk has a positive effect concerning local production. Consumers who care about the origin of the product in general would rather consume organic milk. Furthermore, those that are more frequent consumers of organic
219
milk are influenced by the credence quality associated with organic production process. The credence quality attribute organic process has a positive effect on consumption frequency for organic milk. In contrast, for conventional whole fat milk, organic has a negative effect although it is not significant. The significance of ‘organic’ is important as the study uses self-reported recall of consumption behavior, which might be positively biased as a result of social desirable responses (Verhoef, 2005). Among significant socio-demographics, education has a significant negative effect on all levels for organic milk consumption. A low education decreases the probability of skim milk consumption. Older consumers tend to consume less skim milk as well. There are no significant effects to state for whole fat milk consumption.
5.3 Segmentation of Consumers according to Milk Purchase Behaviour To uncover differences in the consumer groups a hierarchical cluster analysis (Ward linkage, Euclidian distance) was applied. The three cluster solution gave best results. Cluster 1 is characterized by higher income, higher education and with children in the household. This cluster contains consumers who are between 20 and 49 years. Cluster 2 has the highest income and a moderate education level. The consumers in this cluster are 58 years on average. Only 12% have children in the household. Cluster 3 has a higher share of male consumers and is between 69 and 89 years old. The income is lower on average and the education level is rather low. There are almost no children in these households. All clusters prefer whole fat milk, but cluster 1 has a higher consumption of skim milk than cluster 2 and 3 (see table 11). Insert table 11 here. Regarding the use of quality characteristics cluster 1 pays more attention to food safety and functional attributes than cluster 2 and 3. Image attributes seem to be more important for cluster 3. But the elderly seem to pay attention to brands and price, while younger consumers care only about the price. Those between 50 and 67 years use labels such as seals of approval to make purchase decisions. ‘Information seekers’ pay lots of attention towards fat content and labels but little attention towards price and appearance. Compared with the participants in the other clusters the participants in the cluster ‘health awareness’ care more about freshness, health and food safety. They care less about brands and labels. ‘Brand shoppers’ are more interested in brands than the other clusters but have lower interest in nutrition values such as ingredients in general, calories, fat content (see figure 2). Insert figure 2 here. Figure 3 shows differences within the clusters regarding the process quality characteristics. Origin is equally important for all three clusters. Local production is especially important for ‘information seeker’. Moreover, this cluster as well as those with ‘health awareness’ are interested in animal husbandry and organic production. Insert figure 3 here. Figure 4 shows differences within the clusters regarding the environment characteristics. Nutrition attributes are especially important for ‘information seeker’. Furthermore, this cluster pays attention to additional information. The point of sale and a clean source of supply are equally important for all three clusters. Insert figure 4 here. Overall, cluster 1 is most interested in food safety attributes, functional aspects of the product and sensory attributes. Cluster 2 is especially interested in nutritional values, process attributes and all kinds of additional information. Cluster 3 is especially interested in the milk’s shelf life and brands.
220
6.
Final remarks
This research investigated the impact of consumers’ use of certain quality characteristics and socio-demographics on consumption patterns of whole fat milk, skim milk and organic milk. We found a significant positive effect of the quality cue ‘brand’. For milk we can count several brands, retailer as well as manufacturer brands. The significance for brand is even higher for organic milk than for conventional milk. This leads us to conclude that the brand is the cue used by consumers for recognizing the organically produced milk. This shows furthermore that a strong brand could be one method for influencing consumers’ purchase decisions even for fresh almost unprocessed food. The information on fat content has a negative influence on shopping behaviour for organic milk. The results show health conscious consumers chose to buy skim milk. There are different types of fresh milk, e.g. skim milk and non fat milk available. However, there are fewer varieties for organic milk. Thus, there may be opportunities for product line extensions. Finally, we find that the credence quality attribute ‘organic’ has a significant and positive effect on consumption of organic milk. Organic production communicated through a label works as an extrinsic quality cue and can be used for marketing activities. The results of the cluster analysis show that younger consumers are health conscious and this fact could be used for advertisement regarding skim milk and non-fat milk. The older consumers in cluster 2 could be reached by promotional activities in form of leaflets and additional brochures as they are information seekers. The elderly from cluster 3 are concerned with the milks shelf life. Thus, product line extensions such as fresh milk with extended shelf life are the first step to reach this group. Some producers have already been supplying such milks. As this cluster is aware of brands, focus on brand display might be a possibility for grocery store promotions.
7.
References
Alvensleben, R.v., Scheper, U. (1997). The decline of the meat image in Germany – Theoretical considerations. In Alvensleben, R.v., Cramon-Taubadel, S.v., Rohr, A., Scheper, U. and Schleyerbach, K. (eds.): Problems of meat marketing. Department of Agricultural Economics, CAU Kiel, Kiel: working paper 2, 1-7. Bruhn, M. (2004). Lebensmittelqualität – ein Überblick über unterschiedliche Qualitätsbegriffe. In Agrar spectrum, 37, 11-20. Bruhn, M., Grebitus, C., Hümmelink, A. (2005). Die Bedeutung kognitiver Strukturen im Rahmen des Qualitätsurteils der Verbraucher – am Beispiel von Schweinefleisch. In: Hagedorn, K., Nagel, U.J. and Odening, M. (eds.): Umwelt- und Produktqualität im Agrarbereich. Schriften der Gesellschaft für Wirtschafts- und Sozialwissenschaften des Landbaues e.V., Muenster Hiltrup, 40, 91-100. Cardello, A.V. (1995). Food quality: relativity, context and consumer expectations. Food Quality and Preference, 6 (3): 163-170. Caswell, J.A., Bredahl, M.E., Hooker, N.H. (1998). How quality management metasystems are affecting the food industry. Review of Agricultural Economics, 20: 547-557. Darby, M. R., Karni, E. (1973). Free competition and the optimal amount of fraud. Journal of Law and Economics, 16: 67-88. Garvin, D.A. (1984). What Does “Product Quality” Really Mean? Sloan Management Review, 26 (1): 25-43. Gierl, H., Stumpp, S. (1999): Diffusion und Informationswert von Qualitätssignalen. Zeitschrift für Betriebswirtschaft (ZfB), 11 (11): 1241-1268. Grunert, K.G., L. Bredahl, Brunso, K. (2004). Consumer Perception of Meat Quality and Implications for Product Development in the Meat Sector – a Review. Meat Science, 66: 259-272.
221
Grunert, K.G., Larsen, H.H., Madsen, T.K., Baadsgaard, A. (1996): Market Orientation in Food and Agriculture. Massachusetts, USA: Norwell. Kotler, P., Armstrong, G. (1994): Principles of Marketing. London, UK, Prentice-Hall, 6th edition. Nelson, P. (1974). Advertising as information. Journal of Political Economy, 82: 729-754. Northen, J.R. (2000). Quality attributes and quality Cues. Effective communication in the UK meat supply chain. British Food Journal, 102 (3): 230-245. Olson, J.C. (1978). Internal Belief Formation in the Cue Utilization Process. Advances in Consumer Research, 5: 706-713. Olson, J.C., Reynolds, T.J. (1983). Understanding Consumers’ Cognitive Structures: Implications for Advertising Strategy, 77-90. In Percy, L. and Woodside, A.G. (eds.): Advertising and Consumer Psychology. Scholderer, J., Bredahl, L. (2004). Consumer expectation of the quality of pork produced in sustainable outdoor systems. SUSPORKQUAL Deliverable 22: Determination of the weighting of factors influencing attitudes to pork in different countries. Project paper no. 03/04. The Aarhus School of Business, July 2004. Steenkamp, J.-B.E.M. (1990). Conceptual Model of the Quality Perception Process. Journal of Business Research, 21: 309-333. Verhoef, P.C. (2005). Explaining purchases of organic meat by Dutch consumers. European Review of Agricultural Economics, 32 (2): 245-267. ZMP (2002). MAFO Briefe, Aktuelles aus der Agrarmarktforschung. 02/2002, G52795. ZMP (2006a). Milch Marktbilanz 2006. Deutschland, Europäische Union, Weltmarkt. ZMP Zentrale Markt- und Preisberichtsstelle GmbH, Bonn. Rheinbreitbach, Germany: Druckerei Plump KG. ZMP (2006b). Ökomarkt Forum Zentrale Markt- und Preisberichtsstelle (ZMP) – Marktbericht. No. 06, 10th of February 2006, Germany.
222
Tables Table 1.
Subsets of Quality Attributes
Product Attributes Food Nutrition Safety Pathogens Fat Residues Content Hormones Calories Food Fibers additives Sodium Toxins Vitamins GM Minerals Fat/ Cholesterol Physical Contaminants
Sensory Taste Texture Tenderness Juiciness Freshness
Functional Convenience Storage
Image Snob Value Brands Labels
Process Attributes
Environment Attributes Cleanness in the Shop Point of Sale Added Information Recipes Service
Animal Welfare Biotechnology Organic Production Traceability Growth Enhancers Feed
Source: Adapted from Northen (2000), Caswell et al. (1998) and Bruhn et al. (2005). Table 2.
Average of Milk Consumption and Prices in Germany 2003 to 2005
Whole Fat Milk Skim Milk
Mio. litre 682.8 275.3
2003 Ø-price (EUR/l) 0.56 0.53
2004 Mio. Ø-price litre (EUR/l) 630.1 0.56 311.8 0.53
2005 Mio. Ø-price litre (EUR/l) 545.2 0.57 356.2 0.53
Source: ZMP, 2006a: 41.
Table 3.
Per Capita Milk Consumption in Germany 2000 to 2005
Whole Fat Milk Skim Milk
2000 39.2 20.5
2001 40.5 20.3
2002 38.7 21.8
2003 39.0 23.0
2004 36.9 23.8
2005 35.6 25.5
Source: ZMP, 2006a: 41.
Table 4.
Average of Milk Prices in Germany 2004 and 2005 2004
EUR/kg Whole Fat Milk (carton)
Conv. 0.57
Mark Up Conv. to Organic
Organic 0.85
0.28
2005 Conv. 0.57
Organic
Mark Up Conv. to Organic
0.88
Source: Own depiction with data of Verbraucherpreisspiegel ZMP based on GfKHaushaltspanel, ZMP, 2006b.
223
0.31
Table 5.
Structure of the Sample (in %)
Socio-demographics Gender Female Male Age 18 – 34 35 – 49 50 – 64 > 64 Education Low Education Modest Education High Education Very High Education (University Degree) No answer Household Net Income € < 400 400 – 800 800 – 1300 1300 – 1800 1800 – 2300 > 2300 No answer Children in Household Concerned with grocery shopping N total = 260
Table 6.
56 44 40 23 18 19 18 27 35 19 1 7 15 13 19 11 18 17 20 96
Frequency of Consumption Concerning the Different Milks (in %) Daily
Fresh Whole Fat Milk Fresh Skim Milk Organic Milk
28.1 13.1 3.1
5-6 times a week 1.9 2.3 0.0
3-4 times a week 6.9 5.4 1.5
224
1-2 times a week 10.4 5.4 2.3
Less than Never once a week 17.3 35.4 16.2 57.7 15 78.1
Table 7.
Importance of Quality Characteristics on the Purchase of Milk
Categories Product Food Safety Functional
Image
Nutrition
Sensory
Process
Environment
Table 8.
Characteristics Health Hygiene at the cooler Food safety Shelf life Packaging material Packaging size Packaging design Price Labels Brand Fat content Ingredients Calories Freshness Taste Appearance Kind variety Local production Origin Animal husbandry Organic Clean point of sale Point of sale Nutrition information Additional information Overall Quality
48.8 31.2 27.3 86.2 46.2 36.5 18.5 62.7 21.2 20.8 56.9 27.3 18.1 78.1 58.5 20.8 10.4 41.5 36.5 26.2 19.2 73.5 38.5 10.4 8.1 52.7
Definition of Variables
Variables Y Whole Fat Milk (model1) Skim Milk (model2) Organic Milk (model3) X Attributes from table 5 Age Education Income
Household Size Children
Description Dependent Variables Frequency of consumption, such as daily, 5-6 times a week and so on to measure the corresponding latent utilities. Independent Variables Dummy variables equal to one if the consumer marks it as Important / used for purchase of the different milks. Age of the consumer (integer years). (Age squared and log age did not show significant results). Dummy variables for every category (see table 3). Very high education dropped due to multicollinearity. Monthly household net income. Dummy variables for every category (see table 3). 1300-1800 EUR and >2300 EUR dropped due to multicollinearity. Number of persons in the household. Dummy variable equal to one if children in the household.
225
Table 9.
Product Food Safety
a
Estimation Results for Ordered Logit Models for Whole Fat and Skim Milk Consumption
Food safety
Health Packaging Functional design Packaging material Packaging size Shelf life Image Brand Price Labels Nutrition Calories Fat content Ingredients Sensory Appearance Freshness Kind variety Taste Animal Process husbandry Local production Organic Origin Clean POS EnvironMent Additional information Point of sale Nutrition information SocioGender demoAge graphics HH Size Children High EDU Modest EDU Low EDU Y400-800 Y800-1300 Y1800-2300 Overall Quality
Whole fat milk Std. Coef. Err. z-Valuea 0.304 0.397 0.76
Skim milk Std. Coef. Err. z-Valuea 0.509 0.376 1.35
-0.302
0.304
-1
0.505
0.396
1.27
0.360
0.385
0.94
-0.289
0.439
-0.66
-0.186
0.311
-0.6
0.075
0.304
0.25
0.504 -0.296 0.689 -0.126 -0.371 0.026 -0.443 -0.111 0.403 0.750 -0.225 0.382
0.284 0.414 0.361 0.286 0.421 0.406 0.295 0.374 0.344 0.340 0.508 0.293
1.78 -0.71 1.91 -0.44 -0.88 0.06 -1.5 -0.3 1.17 2.21 -0.44 1.3
-0.393 -0.045 0.474 0.245 -0.484 0.048 1.119 -0.590 0.114 0.416 0.539 -0.100
0.319 0.422 0.386 0.304 0.501 0.425 0.352 0.363 0.114 0.416 0.512 0.312
-1.23 -0.11 1.23 0.81 -0.97 0.11 3.18 -1.62 1 1 1.05 -0.32
0.435
0.442
0.98
-0.084
0.427
-0.2
0.595 -0.432 -0.005 0.213
0.310 0.449 0.101 0.329
1.92 -0.96 -0.05 0.65
0.233 0.155 -0.239 0.510
0.338 0.482 0.182 0.397
0.69 0.32 -1.32 1.28
0.234 -0.624
0.465 0.471
0.5 -1.32
0.003 0.041
0.529 0.507
0.01 0.08
-0.513 -0.221 0.002 0.293 0.609 0.194
0.345 0.263 0.009 0.238 0.580 0.397
-1.49 -0.84 0.2 1.23 1.05 0.49
-0.513 0.146 -0.018 -0.150 -0.005 -0.348
0.344 0.300 0.010 0.256 0.609 0.371
-1.49 0.49 -1.84 -0.59 -0.01 -0.94
-0.329 -0.405 0.122 -0.297 -0.457 0.133
0.389 0.489 0.356 0.417 0.417 0.318
-0.85 -0.83 0.34 -0.71 -1.1 0.42
-0.019 -0.893 0.057 -0.220 0.187 0.116
0.396 0.550 0.496 0.506 0.379 0.352
-0.05 -1.62 0.11 -0.44 0.49 0.33
* **
**
**
*** *
*
*
Level of significance: ***p
